Skip to main content

AI Context Setup

GEO Optimizer provides pre-built context files that turn your AI assistant into a GEO expert. Once loaded, the AI understands the Princeton methodology, knows all script flags, can interpret audit results, and can guide content optimization -- without you needing to explain anything.

This page provides detailed setup instructions for each platform, followed by usage examples.

Platform Comparison​

PlatformFileSizeContext DepthSetup
Claude Projectsclaude-project.md~11,700 charsFull methodology + scripts + scoring + botsWeb UI
ChatGPT Custom GPTchatgpt-custom-gpt.md~8,000 charsMethodology + scripts + scoringGPT Builder
ChatGPT Instructionschatgpt-instructions.md~1,500 charsCore concepts + script namesSettings
Cursorcursor.mdcUnlimitedFull context for IDE assistanceFile copy
Windsurfwindsurf.mdUnlimitedFull context for IDE assistanceFile copy
Kirokiro-steering.mdUnlimitedFull context for IDE assistanceFile copy

Claude Projects​

Claude Projects supports the richest context. The claude-project.md file includes the complete methodology, all script references, the scoring algorithm, the AI bot table, and workflow guidance.

Step-by-Step Setup​

  1. Go to claude.ai and open your project (or create a new one)
  2. Click the project name in the sidebar to open project settings
  3. Find the Custom Instructions (or Project Instructions) section
  4. Copy the contents of the context file: 
    cat ~/geo-optimizer-skill/ai-context/claude-project.md | pbcopy   # macOS
    cat ~/geo-optimizer-skill/ai-context/claude-project.md | xclip    # Linux
  5. Paste into the instructions field
  6. Click Save

What Claude Gets​

With the full context loaded, Claude can:

  • Explain each of the 9 Princeton methods with impact percentages
  • Run and interpret geo_audit.py results
  • Generate llms.txt files with proper structure
  • Create JSON-LD schemas for any supported type
  • Configure robots.txt with the correct bot permissions
  • Prioritize optimization recommendations by impact

ChatGPT Custom GPT​

Building a dedicated GPT gives you a reusable GEO assistant with richer context than standard conversations.

Step-by-Step Setup​

  1. Go to chat.openai.com
  2. Navigate to Explore GPTs (sidebar) and click Create
  3. Switch to the Configure tab
  4. In the Name field, enter: GEO Optimizer
  5. In the Description field, enter: Generative Engine Optimization toolkit based on Princeton KDD 2024 research
  6. Copy the context file contents: 
    cat ~/geo-optimizer-skill/ai-context/chatgpt-custom-gpt.md | pbcopy
  7. Paste into the Instructions field
  8. Click Save (choose visibility: Only me, Anyone with a link, or Everyone)
8,000 Character Limit

The GPT Builder instructions field has a strict 8,000 character limit. The chatgpt-custom-gpt.md file is pre-optimized to fit within this limit. Do not add extra text that would exceed it.

What ChatGPT Gets​

The compressed context covers:

  • The 9 GEO methods with priority rankings
  • All three script names with key flags
  • Scoring system overview (sections and point values)
  • AI bot classification (citation vs. training)

It omits the full bot table and some verbose examples to stay within the character limit.

ChatGPT Custom Instructions​

For GEO knowledge in every ChatGPT conversation without creating a separate GPT.

Step-by-Step Setup​

  1. Go to chat.openai.com
  2. Click your profile icon (bottom-left) and select Settings
  3. Navigate to Personalization > Custom Instructions
  4. Copy the context file: 
    cat ~/geo-optimizer-skill/ai-context/chatgpt-instructions.md | pbcopy
  5. Paste into the "How would you like ChatGPT to respond?" field
  6. Toggle Enable for new chats to on
  7. Click Save
Ultra-Compressed Format

At ~1,500 characters, this file contains only the essential concepts: what GEO is, the three script names, and the core methodology. For detailed flag references and scoring breakdowns, use the Custom GPT approach or ask ChatGPT to reference the GitHub repository.

Cursor​

Cursor automatically loads rule files from your project directory as AI context.

Step-by-Step Setup​

# Navigate to your project root
cd /path/to/your-project

# Create the rules directory
mkdir -p .cursor/rules

# Copy the context file
cp ~/geo-optimizer-skill/ai-context/cursor.mdc .cursor/rules/geo-optimizer.mdc

How It Works​

Cursor reads all .mdc files in .cursor/rules/ and includes them as context in every AI interaction within that project. The rules are:

  • Loaded automatically when you open the project
  • Applied to all AI features (Cmd+K, Chat, inline edits)
  • Combined with any other rule files you have
Project-Specific Rules

Place the file in your specific project's .cursor/rules/ directory, not a global location. This way GEO context is only active when working on projects that need it.

Verify It Works​

Open Cursor in your project and ask:

What GEO optimization methods have the highest impact?

If the rules loaded correctly, Cursor will reference the Princeton methods with specific percentages.

Windsurf​

Windsurf uses a similar directory-based rules system.

Step-by-Step Setup​

cd /path/to/your-project

# Create the rules directory
mkdir -p .windsurf/rules

# Copy the context file
cp ~/geo-optimizer-skill/ai-context/windsurf.md .windsurf/rules/geo-optimizer.md

Windsurf loads all markdown files from .windsurf/rules/ and applies them as AI context within the project.

Kiro​

Kiro uses steering files to guide its AI assistant behavior.

Step-by-Step Setup​

cd /path/to/your-project

# Create the steering directory
mkdir -p .kiro/steering

# Copy the context file
cp ~/geo-optimizer-skill/ai-context/kiro-steering.md .kiro/steering/geo-optimizer.md

Kiro reads all markdown files from .kiro/steering/ when providing AI assistance.

Usage Examples​

Once any platform has the GEO context loaded, you can use natural language to leverage the toolkit. Here are practical prompts that work across all platforms.

Run an Audit​

Audit my website https://example.com for GEO compliance.
What's my score and what should I fix first?

The AI will know to use geo_audit.py, explain the scoring, and prioritize recommendations by impact.

Generate llms.txt​

Generate an llms.txt file for my site at https://example.com.
My site name is "Example Corp" and we're a data analytics company.

The AI will use generate_llms_txt.py with the appropriate flags.

Create JSON-LD Schema​

Create a WebSite JSON-LD schema for my Astro site.
Site name: Example Corp, URL: https://example.com

The AI will use schema_injector.py --astro and provide a ready-to-paste snippet.

Optimize Content​

Review this blog post for GEO optimization. Apply the Princeton methods,
focusing on the top 3 (Cite Sources, Statistics, Quotation Addition):

[paste your content]

The AI will analyze the content against all 9 methods and provide specific rewrite suggestions, prioritized by impact.

Configure robots.txt​

Help me set up robots.txt for AI search visibility.
I want to be cited by ChatGPT and Claude but block training crawlers.

The AI will generate a properly configured robots.txt with the correct bot classifications.

Interpret an Audit Report​

Here's my GEO audit JSON output. What are the most impactful changes I should make?

[paste JSON output]

The AI will parse the report, identify the lowest-scoring sections, and provide actionable recommendations ordered by potential impact.

CI/CD Setup​

Help me add a GEO audit step to my GitHub Actions pipeline.
I want it to fail the build if the score drops below 80.

The AI will generate a complete workflow YAML using the JSON output format.

Next: Contributing -- run tests, understand the quality score, and submit pull requests.